'Pacific Island' no longer matches Iceland (IS)

CHANGELOG.md

@@ -3,6 +3,10 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/)
 
+## [5.2.3] - 2024-11-04
+
+- "Pacific Island" no longer matches Iceland.
+
 ## [5.2.2] - 2024-08-23
 
 - Detection for "moldavie" as "MD" in French.

geoconvert/data/countries.py

@@ -48,6 +48,12 @@
     "haute vienne": "FR",  # fr
 }
 
+# We want to make sure pacific island is not registered as IS (Iceland) in German.
+# Other languages doesn't have issues with it and we shouldn't prevent them from retrieving
+# other countries like FJ (Fiji) in "Fiji/Pacific Island"
+special_countries_de = {
+    "pacific island": None,  # en
+}
 
 countries_fr = {
     **special_countries,
@@ -597,6 +603,7 @@
 
 
 countries_de = {
+    **special_countries_de,
     **special_countries,
     **{
         "agypten": "EG",

geoconvert/utils.py

@@ -40,10 +40,12 @@ def safe_string(text):
     'how are you'
     >>> safe_string("l’ocean")
     "l'ocean"
+    >>> safe_string('Fiji/Pacific Island')
+    'fiji pacific island'
     """
     text = remove_accents(text)
-    # Replace "-" and ":" with a whitespace
-    text = re.sub(r"[-:]", " ", text)
+    # Replace "-", ":" and "/" with a whitespace
+    text = re.sub(r"[-:/]", " ", text)
     # Replace weird '
     text = re.sub(r"[ʼ]", "'", text)
     # Only keep word or space characters as well as "_", and "'".

tests/test_countries.py

@@ -65,6 +65,8 @@ class TestCountries:
             ("Solomon Islands", {}, "SB"),  # en
             ("Prince Edward Island", {}, "CA"),  # en
             ("Rhode Island", {}, "US"),  # en
+            ("Pacific island", {}, None),  # en
+            ("Fiji/Pacific island", {}, "FJ"),  # en
             # However, in cases where island is singular instead of plural,
             # there can be confusion.
             ("Solomon Island Nationals", {}, "IS"),  # en
@@ -191,6 +193,7 @@ def test_country_name_to_country_code_en(self, input_data, expected):
             ("   Land der Hinrichtung : Deutschland", "DE"),
             ("Dschibuti-Stadt", "DJ"),
             ("elfenbeinküste und ecuador    ", "CI"),
+            ("Fiji/Pacific island", None),  # en
         ],
     )
     def test_country_name_to_country_code_de(self, input_data, expected):